All Academic, Inc. Research Logo

Info/CitationFAQResearchAll Academic Inc.
Document

Gay Marriage and Civil Unions: The Impact of Network Diversity on Opinion
Unformatted Document Text:  10 inflation factor testing did not find substantial multicollinearity 5 , the correlation between the factors of greatest interest suggested caution. Therefore, an interaction term was created by multiplying the indicator for high network diversity with the indicator for having a gay/lesbian friend. Though interaction terms can induce additional multicollinearity, diagnostic tests still found acceptable VIFs on all variables. Since the primary factor of interest in this study is the diversity of social networks, this allows for the possibility that the impact of general diversity (as measured by the index of race, religion and political views) may vary by the sexual diversity (as measured by having a close gay or lesbian friend). Control and Baseline Categories. Indicator variables were created for gender (1=female); age under 32 or age over 55 (the 25 th and 75 th percentiles, respectively); black or Hispanic ethnicity; Republican or Democrat 6 ; married or partnered (9.7% of survey respondents indicated that they were not married, but living with a partner; however, the survey did not determine whether the partners were same- or opposite-sex); and currently belonged to a fundamentalist religion or currently belonged to a non-fundamentalist religion. Respondents reported their income and education on an ordered, categorical scale that was recoded from 0 (Less than $15,000 for income and less than high school completion for education) to 1 ($200,000 or more for income and post-graduate training for education). The baseline category, therefore, was male, between the ages of 32 and 55, white, no party identification, single, not belonging to any religion, and lowest levels for income and education. Results and Discussion The results of the multinomial logit regression are shown in Table 7. Recall that each regression fits the independent variables to the three paired outcomes for the dependent variable: civil unions versus no legal recognition, full marriage rights versus no legal recognition, and marriage versus civil unions. A preliminary examination provides some useful information. All independent variables with the exception of Democrat party identification display some level of statistical significance. Furthermore, the goodness-of-fit measurements suggest that the model is able to explain some portion of the variance in the observations. Also, there is evidence that the three policy alternatives are, indeed, non-ordered. First, consider the relatively unsurprising results of Republican party identification. Being a Republican, other things being equal, appears to be associated with reduced support for marriage versus nothing, for civil unions versus nothing, and for marriage versus civil unions. Put another way, Republicans 5 A common rule of thumb is that variance inflation factors greater than 10.0 are evidence of multicollinearity; no VIFs in the full model exceeded 2.95. 6 The CID survey provided party identification on the common 7-point scale. Since this issue might be one on which independent voters are divided, respondents were not coded Republican or Democrat if they were “Leaning Republican/Democrat.” Only “Strong Republicans/Democrats” and “Republicans/Democrats” were coded as such.

Authors: Jensen, Micah.
first   previous   Page 10 of 24   next   last



background image
10
inflation factor testing did not find substantial multicollinearity
5
, the correlation between the
factors of greatest interest suggested caution. Therefore, an interaction term was created by
multiplying the indicator for high network diversity with the indicator for having a gay/lesbian
friend. Though interaction terms can induce additional multicollinearity, diagnostic tests still
found acceptable VIFs on all variables.
Since the primary factor of interest in this study is the diversity of social networks, this allows for
the possibility that the impact of general diversity (as measured by the index of race, religion and
political views) may vary by the sexual diversity (as measured by having a close gay or lesbian
friend).
Control and Baseline Categories. Indicator variables were created for gender (1=female); age under 32
or age over 55 (the 25
th
and 75
th
percentiles, respectively); black or Hispanic ethnicity; Republican
or Democrat
6
; married or partnered (9.7% of survey respondents indicated that they were not
married, but living with a partner; however, the survey did not determine whether the partners
were same- or opposite-sex); and currently belonged to a fundamentalist religion or currently
belonged to a non-fundamentalist religion. Respondents reported their income and education on
an ordered, categorical scale that was recoded from 0 (Less than $15,000 for income and less than
high school completion for education) to 1 ($200,000 or more for income and post-graduate
training for education).
The baseline category, therefore, was male, between the ages of 32 and 55, white, no party
identification, single, not belonging to any religion, and lowest levels for income and education.
Results and Discussion
The results of the multinomial logit regression are shown in Table 7. Recall that each regression
fits the independent variables to the three paired outcomes for the dependent variable: civil unions
versus no legal recognition, full marriage rights versus no legal recognition, and marriage versus
civil unions.
A preliminary examination provides some useful information. All independent variables with the
exception of Democrat party identification display some level of statistical significance.
Furthermore, the goodness-of-fit measurements suggest that the model is able to explain some
portion of the variance in the observations.
Also, there is evidence that the three policy alternatives are, indeed, non-ordered. First, consider
the relatively unsurprising results of Republican party identification. Being a Republican, other
things being equal, appears to be associated with reduced support for marriage versus nothing, for
civil unions versus nothing, and for marriage versus civil unions. Put another way, Republicans
5
A common rule of thumb is that variance inflation factors greater than 10.0 are evidence of multicollinearity; no
VIFs in the full model exceeded 2.95.
6
The CID survey provided party identification on the common 7-point scale. Since this issue might be one on which
independent voters are divided, respondents were not coded Republican or Democrat if they were “Leaning
Republican/Democrat.” Only “Strong Republicans/Democrats” and “Republicans/Democrats” were coded as such.


Convention
Need a solution for abstract management? All Academic can help! Contact us today to find out how our system can help your annual meeting.
Submission - Custom fields, multiple submission types, tracks, audio visual, multiple upload formats, automatic conversion to pdf.
Review - Peer Review, Bulk reviewer assignment, bulk emails, ranking, z-score statistics, and multiple worksheets!
Reports - Many standard and custom reports generated while you wait. Print programs with participant indexes, event grids, and more!
Scheduling - Flexible and convenient grid scheduling within rooms and buildings. Conflict checking and advanced filtering.
Communication - Bulk email tools to help your administrators send reminders and responses. Use form letters, a message center, and much more!
Management - Search tools, duplicate people management, editing tools, submission transfers, many tools to manage a variety of conference management headaches!
Click here for more information.

first   previous   Page 10 of 24   next   last

©2012 All Academic, Inc.